I was wondering what the state of Postgres High Availability and how people are implementing them. I understand there are multiple ways of doing this. But I would like to have an experience report where Postgres HA has been configured for HA and automatic failover. I have tried this with streaming replication with repmgr and configured repmgr for failover as well based on response timeout with 2 nodes. I would like to know how or some pointers for best practices where HA for postgres is setup with 3 or more nodes voting for master status and failing over gracefully withtout loss of data. Any latest Postgres HA setup using Pace maker and Corosync and how well it performs.